Habitat use by Siberian warbler species at a stopover site in Far Eastern Russia

ABSTRACT Knowledge of the routes and habitat use of Siberian songbird species during migration is very limited. The goal of our study was to describe the autumnal habitat use of seven Siberian warbler species in the genera Phylloscopus, Acrocephalus, Iduna and Locustella in Far Eastern Russia. A total of 2283 individuals were trapped in mist nets placed within different habitat types between 2012 and 2014 as part of the Amur Bird Project at Muraviovka Park in Far Eastern Russia. We studied the effect of habitat type and vegetation height on the occurrence of each species, and compared our results to published information on habitat use on the breeding grounds. Our results demonstrate that most species exhibit species-specific preferences for habitat type, and that these stopover habitats were similar to habitats used on the breeding grounds.

Most long-distance migrant songbirds migrate nocturnally (Winkler 1989, Liechti & Bruderer 1998 and they make several stopovers in daytime to replenish their energy stores and water reserves for the next bout of flight (Schaub & Jenni 2000, Fuchs et al 2006. Passerine migrants are known to use a range of habitats at stopover sites that is wider than those occupied during breeding season (Bairlein 1983, Bilcke 1984, Chernetsov 2006, although there are still often species-specific habitat preferences (Bairlein 1983). In some species, dominance hierarchies are established, whereby males and adults outcompete females and juveniles for the best habitats (Catry et al 2004, Schaub et al 2004, Tellería & Pérez-Tris 2004, Gyurácz et al 2012, Gyimóthy 2012.
Migrants regularly land in unsuitable habitat after nocturnal movements, often moving within the habitat in the early morning, this manifesting itself as a peak in captures at this time (Brensing 1989, Titov 1999b). Such local movements may be an attempt to establish a defined home range or territory within preferred habitats, and may also be related to increased energy demands after a long flight (Titov 1999a, Chernetsov et al 2004, Chernetsov 2005.
There are many studies dealing with the habitat use and preference of European songbird species, especially warblers (Bairlein 1983, Pambour 1990, Cantos & Tellería 1994, Chernetsov 1998, Schaub & Jenni 2000, Preiszner & Csörgő 2008, Vadász et al 2008a, b, Mérő et al 2016. Previous studies of Siberian songbird species have concentrated on identifying habitat preferences on the breeding grounds , Egorova et al 2009, Alström et al 2011, Bozó 2015 or wintering areas (Round et al 2014, Ayat & Tata 2015, however, and knowledge of habitat use during migration at stopover sites is limited. This lack of knowledge is caused partly by uncertainty over the species' migration routes, but also by a lack of ringing stations in eastern Asia where such studies could be conducted. Information on bird migration at a stopover site in Far Eastern Russia has been collected within the Amur Bird Project since 2011, for species including several Siberian warblers (Heim et al 2012, Bozó & Heim 2015, Sander et al 2017. The variety of vegetation at the ringing station provides an opportunity to examine the habitat use of migratory species.
Using three years' ringing data, our study aimed, first, to identify which habitats were used by seven species of warbler in the genera Phylloscopus, Acrocephalus, Iduna and Locustella at an autumn stopover site in Far Eastern Russia and, second, to compare the stopover habitats with published information on habitat use on the breeding grounds.

Material and methods
The study was carried out, as part of the Amur Bird Project, at Muraviovka Park along the middle stream of the Amur River in the Russian Far East (Heim et al 2012). The study site (49°55'08"N 127°40'20"E) is located 60 km southeast of the city of Blagoveshchensk..
Daily ringing activities were carried out in 2012 (29 August to 16 November), 2013 (25 July to 23 October) and 2014 (25 July to 29 September). Twenty one mist nets (216 m in total length) were set up in a variety of habitats (with the dominant plant species in parentheses): deciduous forest (Quercus mongolica, Lespedeza bicolor), pine forest (Pinus sylvestris), homogeneous reed beds (Phragmites australis), shrub (Salix sp., other unidentified bushes), and sedges and grassy swamps interspersed with willows (Carex sp., Poa sp.). For the analysis we assigned each mist net to one of the following habitat types: deciduous forest (42 m of mist nets), pine forest (24 m), reeds (24 m), shrub (42 m) and swamp (84 m). Mist nets were checked every full hour from sunrise to sunset. Rings were supplied by the Moscow Ringing Centre. Species identification was based on Svensson (1992) and Brazil (2009).
We also studied the potential effects of vegetation height on the birds' use of habitat. In 2014 we mapped the vegetation types and recorded the average height of the vegetation around the nets.
Linear models were fitted for each species, with the number of birds trapped at each mist-net site (per netm per day) as response variable, and habitat type and vegetation height as factors. Significant variables were selected with the help of backward stepwise model selection (Crawley 2013) using the likelihood-ratio test (P < 0.05). Data analysis was performed in Microsoft Excel 2017 and R version 3.4.2 (R Development Core Team 2017).

Results
No relationships between the number of birds (per netm per day) and vegetation type or height were found for Thick-billed Warbler or Radde's Warbler. Significant relationships between capture rates of birds and both habitat type and vegetation height were found for Yellow-browed Warbler: this species preferred higher vegetation and the 'shrub' habitat type. Significant relationships with habitat type but not with vegetation height were found for Dusky Warbler, Pallas's Leaf Warbler, Black-browed Reed Warbler and Pallas's Grasshopper Warbler. Whereas the latter two species preferred reeds, Dusky Warbler preferred swamp and Pallas's Leaf Warblers both types of forest (deciduous and pine) as well as shrub (Table 2).

Discussion
Our results show that most of the species studied have specific patterns of habitat use, which are likely to be related to food preferences and foraging methods.
The Thick-billed Warbler breeds in dense thickets, bushes, tall grass, the edges and clearings of forests and other wetlands (Dyrcz 2017a). Our results revealed that neither habitat type nor vegetation height can explain the pattern of occurrence of the species in the study area. The same was found for Radde's Warbler, which breeds in taiga forests with dense undergrowth (Clement 2017a). This result may stem from these species using a variety of habitat types on migration.
The Black-browed Reed Warbler breeds in sedges with clumps of willows and other bushes, tall grass and reed beds (Dyrcz 2017b) and was also found to use reed beds during migration stopovers at our site. Our results show that this species is more associated with reed beds than the other warblers studied. This shows that the wetlands are very important for the species both in breeding and migratory periods.
Pallas's Grasshopper Warbler breeds in tall grass with tussocks and riverbanks, valley meadows, moist taiga woodland and shrubby steppes (Pearson 2017) and was found to use reed beds during stopover on migration. Therefore, we can say that the species uses wetland habitats both on migration and when breeding.
The Yellow-browed Warbler and Pallas's Leaf Warbler breed in taiga forests and feed within the canopy (Alström et al 2018. Both species feed on insects, which are hunted in forest canopy. However, significant correlation between species occurrence and vegetation height was found only for Yellow-browed Warbler. When feeding, both species often hover and snatch insects from the edges of leaves; this foraging method can be used easily in the canopy but would be very difficult in denser vegetation (Forstmeier & Keßler 2001). Pallas's Leaf Warbler was also found in large numbers in the pine forest on migration, indicating that they use the same habitat in both breeding and migratory periods. The Dusky Warbler breeds in bushes, taiga forests, and wetlands bordering bushes, and feeds in lower vegetation (Clement 2017b). Our results indicate that on migration this species prefers sedges and grassy habitat, unlike the habitats used during the breeding season.
Usually migratory passerines seem to use wider habitat preferences during stopovers than in the breeding period, behaving as generalists in their habitat selection (Bairlein 1983, Vega Rivera et al 2003, Chernetsov 2006, Preiszner & Csörgő 2008. Among other species, Wrens Troglodytes troglodytes (Hawthorn 1971(Hawthorn , 1975 and Blue Tits Cyanistes caeruleus (Gyurácz et al 2012) use a variety of habitats during migration. In contrast, six of the seven species in our study use the same or similar habitats as in the breeding season, highlighting the importance of protecting a wide variety of habitatsas different species use different habitats. Three of the seven species associate with wetlands on migration as well as when breeding but, unfortunately, wetlands are among the most endangered habitats of eastern Asia. If the extent of these habitats (reed beds, marshes and other wetlands) were to decrease drastically in future, the population of these species might decrease as well. A recent study estimated that Yellow-browed Warblers need to stop at least six times during autumn migration before reaching their wintering grounds (Sander et al 2017), highlighting the importance of abundant habitat availability all along the route.
In eastern Asia the number of sites with natural habitats (particularly wetlands and woodlands) has been drastically reduced and illegal trade and hunting of birds poses a serious threat even to once-common species (Kamp et al 2015, Edenius et al 2016. Thus, knowledge of year-round habitat use of migratory species is crucial to their future conservation. volunteers for the Amur Bird Project who took part in the fieldwork. This study was supported by the Förderkreis Allgemeine Naturkunde Biologie FAN(B) eV, the German Ornithologists' Society (DO-G eV) and the Oriental Bird Club. LB would like to thank the local government of Kevermes, the contractors of Kevermes and the Campus Hungary Studentship that sponsored his research. Many thanks also to Tamás Sipos for his help with English translation and correction. The comments of two anonymous reviewers greatly improved an earlier version of this manuscript.